| Ontology is a philosophical concept originally, and it describes the natures of objects. In the fields of computer science, ontology is a formal specification of a shared conceptualization. Because there is not a common criterion for building ontology, a lot of heterogeneous ontologies are constructed in the same domain.Ontology mapping is an effective method to realize the interoperation of heterogeneous ontology, the key of ontology mapping is to calculate similarities between concepts. After a comprehensive study of the basic theory of ontology and the knowledge of ontology mapping, a new multi-strategy ontology mapping method (MSOM) is put forward, and the concept similarity is computed from the aspects of concept name, concept attribute and concept relationship. What's more, the candidate set and information gain are brought to improve the accuracy of similarity and to simplify the calculation process. The following are the points of this paper:(1). Introduces the origin and the evolvement of ontology, introduces the evolvement of the describe language of ontology.(2). A detailed studying on the basic theory of ontology mapping and several well-known mapping systems currently.(3). Aimed at the problem of that the computation of similarities is not perfect, that the computation is too complex and that the computation is unilateral. So a new ontology mapping method is put forward, and the concept similarity is computed from the aspects of concept name, concept attribute and concept relationship. What's more, the candidate set and information gain are brought to improve the accuracy of similarity and to simplify the calculation process.(4). With the new method of computing similarity, an ontology mapping framework MSOMF (Multi-Strategy Ontology Mapping Framework) is designed.(5). Extensive simulations show that the approach is an effective algorithm and better than present others.At last, the dissertation concludes by summarizing the research and indicating its future way. |
PDF Full Text Request